The End of My Addiction

family member interview; (3) determination of alcohol concentration in blood and saliva by QED (Enzymatics Inc., Horsham, UK). Cumulative abstinent duration (CAD), defined as the total number of days of abstinence, was also calculated in both the baclofen and placebo groups. Further, main biological markers of alcohol abuse [aspartate aminotransferase (AST), alanine aminotransferase (ALT), [.gamma]-glutamyltranspeptidase (GGT) and mean cell volume (MCV)] were determined at the start (T0) and at the end (T4) of the study. Finally, possible changes in state anxiety and current depression were assessed by means of the State and Trait Inventory test, Y1 axes (Spielberg et al ., 1983), and Zung Self-rating Depression Scale (Zung et al ., 1965), respectively.
    At drug discontinuation, the presence of possible side-effects due to drug suspension was recorded on a weekly basis for the first 4 weeks.
    Statistical evaluation of patients’ age, years of addiction and CAD in the baclofen and placebo groups was performed by the Mann–Whitney test. The number of drop-outs and of patients maintaining abstinence in the two groups were compared using the Fisher exact test for a 2 × 2 table [treatment (baclofen; placebo) × drop-out (presence; absence) or treatment (baclofen; placebo) × abstinence (presence; absence)]. The numbers of patients maintaining abstinence and CAD were analysed with the intention-to-treat principles (see Lehert, 1993), i.e. entering into the analysis any randomized patient, including drop-outs. In this analysis, it was assumed that all patients who terminated treatment before the end of the study were abstinence failures and CAD was calculated on the data available at the time of the last weekly visit. Analysis of the effect of baclofen on daily drinks, OCDS scales, scales of state anxiety and depression, and main biological markers of alcohol misuse was performed by the two-way (treatment × time) analysis of covariance (ANCOVA) with repeated measures on the time factor, and baseline data as covariance.
    Results
    No statistically significant difference in mean age and mean years of addiction was found between the two groups ( P > 0.05, Mann–Whitney test).
    A schematic diagram on recruitment, group allocation, treatment retention and success in achieving and maintaining complete abstinence is presented in Figure 1. Although statistical significance was not reached ( P = 0.06, Fisher’s exact test), the number of drop-outs was lower in the baclofen than in the placebo group; indeed, three subjects in the baclofen group (corresponding to 15.0%) and eight subjects in the placebo group (42.1%) dropped out and were excluded from further statistical analyses.

    Fig. 1. Diagram on recruitment, group allocation, treatment retention and success in achieving and maintaining complete abstinence.
    A significantly higher number of patients who achieved and maintained abstinence throughout the experimental period was found in the group of patients treated with baclofen (14 out of 20, corresponding to 70.0%) compared to subjects treated with placebo (four out of 19, or 21.1%) ( P < 0.005; Fisher’s exact test). CAD was [.similar]3-fold higher in baclofen-than placebo-treated patients [19.6 ± 2.6 and 6.3 ± 2.4 (mean ± SEM), respectively; P < 0.005, Mann–Whitney test].
    Figure 2 shows daily alcohol intake in the two groups of patients at the different observation times of the study. ANCOVA revealed a significant effect of treatment on alcohol intake [ F treatment (1,78) = 10.71, P < 0.005; F time (3,78) = 1.38, P > 0.05]. In the baclofen group, the mean number of daily drinks was virtually completely suppressed within the first week of the treatment, being reduced from ~18 (value at T0) to < 0.5 (values at T1–T4); in the placebo group, the daily drinks were reduced from approximately a mean number of 10 (T0) to 3.5–4.5 (T1–T4).

    Fig. 2. Number of daily drinks in baclofen and placebo groups at T0 (baseline) and over the four weekly visits (T1–T4). Each value is the mean ± SEM of 17 patients in the baclofen group and 11 in the placebo group.
    Figure 3 (top panel) shows the craving score in the two groups of patients at the different observation times. ANCOVA showed a significant effect of both treatment and time on total OCDS score [ F treatment (1,78) = 5.65, P < 0.05; F time (3,78) = 10.30, P < 0.00005]. From T1 to T4, the score in the baclofen group was constantly lower than that monitored in the